Cardiovascular effects of cannabinoids

Pharmacology & Therapeutics 95 (2002) 191 – 202

Cardiovascular effects of cannabinoids Michael D. Randall*, David Harris, David A. Kendall, Vera Ralevic School of Biomedical Sciences, University of Nottingham Medical School, Queen’s Medical Centre, Nottingham NG7 2UH, UK

Abstract The prototypic endocannabinoid, anandamide, and synthetic analogues have been shown to elicit pressor and depressor effects, bradycardia, vasorelaxation, and inhibition of neurotransmission in the central and peripheral nervous systems. Cannabinoid-mediated inhibition of neurotransmission is mediated by inhibition of voltage-gated Ca2 + channels and adenylyl cyclase and activation of inwardly rectifying K + channels. The precise mechanisms underlying the vasorelaxant actions of cannabinoids are currently unclear, but might involve both receptor-dependent and -independent and endothelium-dependent and -independent pathways. Mechanisms proposed have included the release of endothelial autacoids, activation of myoendothelial gap junctions, activation of the Na + pump, activation of K + channels, inhibition of Ca2 + channels, and activation of vanilloid receptors, leading to the release of sensory neurotransmitters. Pathophysiologically, the vasodilator actions of endocannabinoids have been implicated in the hypotension associated with both septic and haemorrhagic shock, but their physiological significance remains to be determined. D 2002 Elsevier Science Inc. All rights reserved. Keywords: Endocannabinoids; Anandamide; Vasorelaxation; Endothelium; Gap junctions Abbreviations: 2-AG, 2-arachidonoylglycerol; CGRP, calcitonin gene-related peptide; DRG, dorsal root ganglion; EDHF, endothelium-derived hyperpolarising factor; 18a-GA, 18a-glycyrrhetinic acid; HUVEC, human umbilical vein endothelial cell; NO, nitric oxide; RVLM, rostral ventrolateral medulla; VR, vanilloid receptor.

Contents 1. 2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . Cardiovascular effects . . . . . . . . . . . . . . . . . . . . 2.1. Actions of cannabinoids in vivo . . . . . . . . . . . 2.2. Uptake and metabolism . . . . . . . . . . . . . . . 3. Actions of cannabinoids in vitro . . . . . . . . . . . . . . 3.1. Vasorelaxation . . . . . . . . . . . . . . . . . . . . 3.2. Vasoconstriction . . . . . . . . . . . . . . . . . . . 3.3. Modulation of sympathetic neurotransmission . . . . 3.4. Modulation of sensory neurotransmission . . . . . . 4. Sites of action . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Vascular cannabinoid receptors . . . . . . . . . . . 4.2. Receptor-independent actions of cannabinoids. . . . 5. Where are endocannabinoids produced in the vasculature? . 6. Endocannabinoids and pathophysiology . . . . . . . . . . 7. Concluding remarks . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . .

* Corresponding author. Tel.: +44-115-9709484; fax: +44-115-9709259. E-mail address: [email protected] (M.D. Randall). 0163-7258/02/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved. PII: S 0 1 6 3 - 7 2 5 8 ( 0 2 ) 0 0 2 5 8 - 9

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

192 192 192 193 194 194 195 196 196 198 198 198 199 200 200 200 200

192

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

1. Introduction Whilst cannabis has been widely used for social and medicinal purposes over several thousands of years, the recognition that there are endogenous cannabinoids has renewed the interest in the pharmacology of cannabinoids. Most of the recent research has focussed on the actions of cannabinoids on the nervous system, but it is now clear that endogenous cannabinoids exert vascular effects, which may be of physiological or pathophysiological significance. The aim of this review is to examine the cardiovascular effects of endogenous cannabinoids and to consider them alongside the effects of exogenous cannabinomimetics.

2. Cardiovascular effects 2.1. Actions of cannabinoids in vivo The in vivo cardiovascular effects of cannabinoids are complex, with both increases and decreases in blood pressure being reported (Stark & Dews, 1980; Dewey, 1986). Cannabinoids can potentially modulate autonomic outflow in both the central and peripheral nervous systems, as well as have direct effects on the vasculature. However, their peripheral actions appear to predominate in cardiovascular control, at least upon systemic administration at the doses used by most investigators. Moreover, their actions will necessarily be complicated by their rapid metabolism, which may liberate other vasoactive substances and their precursors (see Section 2.2). An underlying theme is that systemically administered cannabinoids cause hypotension and bradycardia by peripheral prejunctional inhibition of sympathetic outflow and increased vagal activity, respectively. With respect to the endogenous cannabinoid anandamide, it has been shown to cause bradycardia (with brief secondary hypotension), then a transient pressor effect, which is followed by a delayed, but maintained, depressor action in anaesthetised rats (Varga et al., 1995, 1996; Lake et al., 1997). The initial bradycardia and associated hypotension is believed to be vagally mediated, as it is abolished by atropine treatment or cervical vagotomy (Varga et al., 1995, 1996). Similarly, in mice, both anandamide and synthetic cannabinoid receptor agonists cause biphasic hypotension (a sharp depressor response, followed by a more sustained hypotensive phase, but without a pressor component), which is thought to be entirely CB1 receptor-mediated, as the responses are absent in CB1 receptor knockout mice (Ledent et al., 1999). In the rat, the second depressor effect, which follows the transient pressor phase, is believed to be mediated by CB1 receptor prejunctional inhibition of sympathetic outflow in the periphery, as the effect is attenuated by cervical spinal transection, a-adrenoceptor, and cannabinoid receptor antagonists (Varga et al., 1995, 1996; Lake et al., 1997). The rostral ventrolateral medulla (RVLM) is an area of the

brainstem considered to be of fundamental importance in the mediation of reflex control of the cardiovascular system, and the activity of its sympathetic premotor neurones parallels sympathetic outflow in the periphery. However, the second depressor response to anandamide does not appear to involve centrally induced sympathoinhibition (although its magnitude is dependent on the level of sympathetic outflow), as there was only an increase in the activity of barosensitive RVLM neurones in baroreceptorintact rats, and activity remained unchanged in barodenervated rats (Varga et al., 1996). The sympathoinhibitory action was greater in spontaneously hypertensive rats compared with normotensive controls, perhaps reflecting the higher level of sympathetic tone in the former (Lake et al., 1997). Similarly, recent work by Niederhoffer and Szabo (1999), carried out on pithed rabbits, in which sympathetic tone was evoked by continuous electrical stimulation, has demonstrated that intravenous injection of CB1 receptor agonists (CP-55940 and WIN-55212-2) causes prejunctional inhibition of sympathetic activity, leading to hypotension. Although the second depressor effect of systemic cannabinoid administration does not appear to involve central cardiovascular control centres, the nucleus tractus solitarius (which has pathways to the RVLM and other brainstem centres important in controlling sympathetic outflow) is a possible central site for cannabinoid regulation of cardiovascular function as cannabinoids (CP-55,940 and WIN-55212-2) inhibited the activity of neurones in this region in rat brain slices (Himmi et al., 1998), and immunoreactivity for CB1 receptors was detected in many of the fibres (Tsou et al., 1998). It should be noted that many of the above studies do not exclude the possibility that the second hypotensive effect of cannabinoids additionally involves direct vasorelaxant effects on the blood vessels, as shown extensively in blood vessels in vitro. Indeed, not all investigators agree that there is an involvement of the autonomic nervous system in cannabinoid-mediated hypotension and bradycardia. Vidrio et al. (1996) reported that the cannabinoid agonist HU210, on administration to both conscious and anaesthetised rats, caused prolonged hypotension and bradycardia, but this was independent of the sympathetic nervous system (it was unaffected by chemical sympathectomy), and vagotomy inhibited only the late stage of bradycardia. There was also a lack of vasopressor response to HU210, which may have contributed to its potency as a vasodepressor. The pressor effects of cannabinoids observed on intravenous administration appear to be independent of central control mechanisms and of cannabinoid CB1 receptors. In rats, the pressor response to anandamide was not blocked by a-adrenoceptor blockade, cervical spinal cord transection, or the CB1 receptor antagonist SR141716A (Varga et al., 1996; Lake et al., 1997). The mechanism is unclear, but direct contractile actions on the vascular smooth muscle are possible. Although a central involvement in the pressor response was ruled out, centrally administered cannabinoids

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

can increase the activity of sympathoexcitatory neurones in cardiovascular regulatory centres. Therefore, the potential exists that they may increase blood pressure via actions in the CNS. Injection into the cisterna cerebellomedullaris of CB1 receptor agonists (WIN-55212-2 and CP-55940) in conscious rabbits caused an increase in sympathetic nerve activity, plasma noradrenaline concentrations, and blood pressure, and this action was blocked by the CB1 receptor antagonist SR141716A (Niederhoffer & Szabo, 1999). Currently known mechanisms of cannabinoid receptor signalling would suggest that they principally act via inhibition of neurotransmitter release, indicating that the central sympathoexcitatory effects of cannabinoids may be mediated by disinhibition. The majority of the above-mentioned cardiovascular studies were carried out in anaesthetised animals. A significantly different profile of haemodynamic responses to systemic cannabinoid administration is observed in conscious animals, in that the pronounced depressor response to cannabinoids is generally weak or absent. In the conscious rat, Stein et al. (1996) reported that anandamide caused bradycardia, with a transient hypotensive effect, followed by a longer pressor phase, and only at the higher doses was there delayed hypotension. It seems likely that the greater pressor effect obscures the hypotension. Similarly, in the conscious rat, intravenous administration of anandamide and the synthetic cannabinoid WIN-55212-2 caused a pressor effect (Ralevic et al., 2000a; Gardiner et al., 2001). Only following ganglion blockade and in the presence of angiotensin II and vasopressin receptor antagonists was the pressor response to WIN-55212-2 reduced and a delayed hypotensive response uncovered. These findings are consistent with the pressor effects of cannabinoids being mediated via neurohormonal pressor systems, and do not support cannabinoids as having major vasodilator actions in conscious rats in vivo. In humans, acute administration of cannabinoids is associated with tachycardia and a small pressor effect, whereas long-term use is associated with hypotension and bradycardia (Benowitz & Jones, 1975; Benowitz et al., 1979).

193

with arachidonic acid metabolites contributing at more distant sites. Consistent with the apparent local metabolism of endocannabinoids, Calignano et al. (1997) reported that the hypotensive response to anandamide in guinea-pigs (which occurs independently of the autonomic nervous system and is mediated via CB1 receptors) was enhanced by inhibition of carrier-mediated anandamide uptake with AM404. Furthermore, in the rat isolated mesenteric arterial bed, inhibition of the cannabinoid transporter with AM404 or bromocresol green augmented vasorelaxation to anandamide (Fig. 1) (Harris et al., 1998). In contrast, Chaytor et al. (1999) reported that AM404 attenuated relaxations to anandamide in the rabbit mesenteric artery. An implication of this latter finding is that the cannabinoid transporter is involved in allowing anandamide access to intracellular sites, where it may exert vascular effects. However, the actions of AM404 are not solely confined to inhibiting the uptake of endocannabinoids, as AM404, which is structurally related to anandamide and capsaicin, is a potent agonist at endogenous vanilloid receptors (VRs) on sensory nerves and at the recombinant VR (VR1) (Zygmunt et al., 2000a; Ralevic & Kendall, 2001), indicating that caution must be applied in interpreting its actions. An anandamide transporter has been identified on human umbilical vein endothelial cells (HUVECs), and is regulated via nitric oxide (NO) (Maccarrone et al., 2000). Specifically, in HUVECs, there is facilitated uptake of anandamide, which is increased in the presence of NO, indicating that NO may regulate the vascular effects of anandamide. Not only do endothelial cells have specific transporters for anandamide, but they may also possess the amidohydrolase responsible for its metabolism. Evidence for this

2.2. Uptake and metabolism The cardiovascular actions of endocannabinoids are complicated by their rapid metabolism. In this regard, 2arachidonoylglycerol (2-AG), which is believed to be a natural ligand for both CB1 and CB2 receptors, is particularly unstable. Indeed, administration of 2-AG to mice was shown to cause CB1 receptor-independent effects that were mediated via arachidonic acid metabolites (Jarai et al., 2000). However, when a metabolically stable analogue of 2-AG was applied, the hypotensive effects were mediated via CB1 receptors. The rapid metabolism of 2-AG, therefore, might point to the actions of endocannabinoids, in particular 2-AG, being localised to their site of production,

Fig. 1. Anandamide-induced vasorelaxation in the isolated perfused mesenteric arterial bed of the rat. The graph shows that in the presence of either AM404 (3 mM) or bromocresol green (BCG, 30 mM), both of which block the reuptake of anandamide, the potency of anandamide is increased. Addition of either AM404 or BCG produced reductions in tone (AM404: from 161 ± 6 mm Hg to 87.4 ± 5.7 mm Hg; BCG: 157 ± 6 mm Hg to 70.6 ± 6.0 mm Hg), but neither agent affected vasorelaxation induced by the K + -channel activator levcromakalim. Data are shown as mean ± S.E.

194

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

comes from the observation that anandamide-induced NO release in the human saphenous vein and the internal mammary artery is augmented following inhibition of the amidohydrolase (Bilfinger et al., 1998).

3. Actions of cannabinoids in vitro 3.1. Vasorelaxation The first report that anandamide was a vasodilator came from Ellis et al. (1995), who demonstrated that anandamide caused cerebrovascular vasodilatation in the rat that was sensitive to indomethacin. This led to the suggestion that endocannabinoids cause relaxation through the stimulation of the metabolism of arachidonic acid. Furthermore, D9tetrahydrocannabinol was also found to act in this way. These effects could represent direct or indirect effects of cannabinoids on prostanoid production from vascular or extravascular tissue. Following the identification of anandamide as a vasorelaxant, we also demonstrated similar actions in the rat isolated mesenteric and coronary vasculatures (Randall et al., 1996; Randall & Kendall, 1997). However, in mesenteric arterial vessels (Randall et al., 1996, 1997; Plane et al., 1997; White & Hiley, 1997) and the coronary vasculature (Randall & Kendall, 1997), anandamide-induced relaxation was insensitive to cyclo-oxygenase inhibitors and also endothelial denudation (Fig. 2) (Randall et al., 1996; White & Hiley, 1997). Indeed, the only other report of sensitivity to cyclooxygenase inhibition comes from Fleming et al. (1999), who found that diclofenac abolished anandamide-mediated vasorelaxation in rat mesenteric arterial vessels. Thus, the weight of evidence indicates that an involvement of prostanoids in anandamide-induced vasorelaxation may be restricted to certain blood vessels. The relaxant effects of anandamide show tissue selectivity, as it does not relax certain conduit vessels, such as rat carotid arteries (Holland et al., 1999) or the rat aorta (Darker et al., 1998). However, anandamide does relax conduit vessels, such as the rat hepatic and guinea-pig basilar arteries (Zygmunt et al., 1999) and the bovine coronary artery (Pratt et al., 1998). Therefore, it may be that there are regional and species differences in the actions of endocannabinoids. Anandamide has been shown to act via the release of endothelium-derived NO in the rat kidney (Deutsch et al., 1997). A range of human blood vessels and the right atrium have also been shown to release NO in response to anandamide (Bilfinger et al., 1998). However, in many instances (see Randall et al., 1996; White & Hiley, 1997; Jarai et al., 1999), vasorelaxant responses to anandamide are insensitive to inhibition of NO synthase (Fig. 2). In HUVECs, Maccarrone et al. (2000) reported that anandamide and the CB agonist HU210 both cause an up-regulation of the expression and activity of the inducible NO synthase. This is unlikely to contribute toward the acute vascular effects of

Fig. 2. In the isolated perfused mesenteric arterial bed of the rat, anandamide-induced vasorelaxation is unaffected by (a) inhibition of NO synthesis [with 300-mM NG-nitro-L-arginine methyl ester (L-NAME)], inhibition of cyclo-oxygenase [with 10-mM indomethacin (indo)], or (b) removal of the endothelium (-endothelium). Data are shown as mean ± S.E.

anandamide, although a role in pathophysiological situations can be envisaged. Most studies have shown that the vasorelaxant responses to anandamide are endothelium-independent (Randall et al., 1996; White & Hiley, 1997) or only partly endotheliumdependent (Chaytor et al., 1999). However, in the bovine coronary artery, anandamide induces relaxations that are strictly endothelium-dependent (Pratt et al., 1998). This is explained by the endothelial cells metabolising exogenous anandamide, via a cytochrome P450 mono-oxygenase, to vasoactive metabolites. We originally proposed that anandamide might act via a hyperpolarising mechanism (Randall et al., 1996). This was based on the finding that anandamide-induced vasorelaxation was abolished by raised extracellular K + (Randall et al., 1996; White & Hiley, 1997). Indeed, partly based on this observation, we proposed that anandamide might represent the elusive endothelium-derived hyperpolarising factor (EDHF) (for reviews, see Garland et al., 1995; Mombouli & Vanhoutte, 1997). This initiated a substantial research effort to examine this proposal. The electrophysiological actions of anandamide on vascular smooth muscle were first

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

characterised by Plane et al. (1997). In their study, they found that anandamide caused vascular smooth muscle hyperpolarisation or repolarisation, but this was independent of cannabinoid CB1 receptors. Vascular smooth muscle hyperpolarisation was also found to be endothelium-dependent (Chataigneau et al., 1998; Zygmunt et al., 1997), with the implication that anandamide acted via the release of EDHF. The latter study also provided evidence that anandamide acted via inhibition of Ca2 + mobilisation in vascular smooth muscle cells, without direct effects on K + conductance, a finding that was later supported by others (White & Hiley, 1998). Consistent with vascular smooth muscle hyperpolarisation, anandamide induced vasorelaxation in the rat mesenteric arterial bed that was sensitive to nonspecific K + channel blockers, including cytochrome P450 inhibitors (Randall et al., 1997). There is conflicting evidence for an involvement of Ca2 + -activated K + channels in anandamideinduced vasorelaxation. In isolated mesenteric arterial segments, the relaxation response to anandamide was blocked by selective inhibitors of large conductance Ca2 + -activated K + channels (charybdotoxin and iberiotoxin) (Plane et al., 1997). Furthermore, in similar mesenteric vessels, the anandamideinduced relaxation was insensitive to the combination of charybdotoxin and apamin (White & Hiley, 1997). In contrast, in the perfused mesenteric arterial bed, the combination of charybdotoxin and apamin abolished the relaxation response to anandamide, although neither agent alone affected this response (Randall & Kendall, 1998). In human brain endothelial cells, 2-AG has been shown to reverse endothelin-1-stimulated Ca2 + mobilisation via the activation of Ca2 + -activated K + channels (Chen et al., 2000). In the guinea-pig carotid artery, the anandamide-induced hyperpolarisation, which was insensitive to charybdotoxin plus apamin, was blocked by the ATP-sensitive K + channel inhibitor glibenclamide (Chataigneau et al., 1998). However, glibenclamide does not affect anandamide-induced relaxation in the rat mesentery (Randall et al., 1997; White & Hiley, 1997). In conclusion some, but not all, studies point to the involvement of K + channels, at some stage, in the vasorelaxant actions of anandamide. The finding that anandamide caused endotheliumdependent hyperpolarisation coupled with sensitivity, in some cases, to K + channel inhibitors, raised the possibility that anandamide might act in an endothelium-dependent manner via the release of EDHF. However, in this respect, initial studies indicated that vasorelaxation was preserved following removal of the endothelium (Randall et al., 1996; White & Hiley, 1997) (Fig. 2). In contrast, in rabbit mesenteric vessels, anandamide acts partly in an endothelium-dependent manner following uptake into the endothelial cells (Chaytor et al., 1999). The resulting endothelialdependent component to vasorelaxation was blocked by selective peptide gap junction inhibitors and 18a-glycyrrhetinic acid (18a-GA), implicating the involvement of gap junctions. Indeed, it is now recognised that myoendothelial

195

gap junctions might play a substantial role in the phenomenon of endothelium-dependent hyperpolarisation ascribed to EDHF (Chaytor et al., 1998; Harris et al., 2000). The proposal that anandamide might be an EDHF was based, in part, on the finding that EDHF-mediated responses were sensitive to the CB1 receptor antagonist SR141716A (Randall et al., 1996). In this respect, Chaytor et al. (1999) reported that SR141716A was a potent inhibitor of myoendothelial gap junctions, which would account for the actions of this antagonist against EDHF-mediated responses. In rat mesenteric vessels, Wagner et al. (1999) identified a small endothelial component of relaxation to anandamide that was SR141716A-sensitive, but not mediated by CB1 receptors. From this finding, they proposed that there is a novel endothelial cannabinoid receptor. An alternative explanation for the sensitivity to SR141716A, but independence of CB1, perhaps could be additional effects of the antagonist, for example, inhibition of gap junctional communication (Chaytor et al., 1999). Further work in this area has indicated that a neurobehaviourally inactive cannabinoid, abnormal cannabidiol, causes SR141716A-sensitive mesenteric vasodilatation that is also blocked by cannabidiol (Jarai et al., 1999). From these findings, it was proposed that cannabidiol was an antagonist of this novel endothelial cannabinoid receptor, which is coupled to EDHF release. The possibility that anandamide-induced vasorelaxation might involve EDHF release and/or gap junctional communication led us to investigate further the pharmacology of this vasorelaxation response. In this respect, the endothelium-independent vasorelaxation was sensitive to gap junction inhibitors (18a-GA and ouabain) that also block the Na + pump, but was unaffected by agents that are selective for gap junctions (carbenoxolone and palmitoleic acid). This raises the possibility that the Na + pump, at some stage, may be involved in the vasorelaxation response to anandamide, independently of any contribution of EDHF (Fig. 3). Endocannabinoids have been shown to inhibit vascular smooth muscle Ca2 + channels (Gebremedhin et al., 1999). Specifically, in feline cerebral vessels, it was shown that endocannabinoids and synthetic cannabinoid agonists act via G-protein-coupled CB1 receptors to cause inhibition of voltage-sensitive Ca2 + channels, leading to vasodilatation. This action was proposed to contribute toward vasodilatation in cerebral hypoxia, which was associated with the release of endocannabinoids. 3.2. Vasoconstriction Contractions mediated by anandamide and synthetic cannabinoids have been reported in rat isolated small mesenteric arteries (White & Hiley, 1998). The responses were small, and it was suggested that only low levels of Ca2 + were released from intracellular stores, which were insufficient to stimulate extracellular Ca2 + entry. In the anaesthetised rat, anandamide, despite causing widespread vasodilatation, recently has been shown to cause SR141716A-insensitive

196

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

Fig. 3. Anandamide-induced vasorelaxation in the isolated perfused mesenteric arterial bed of the rat in the presence of the combined gap junction and Na + pump inhibitors (a) 1-mM ouabain and (b) 100-mM 18a-GA and pure gap junction inhibitors (c, 50-mM palmitoleic acid; d, 100-mM carbenoxolone). Ouabain alone reduced established tone from 136 ± 7 mm Hg to 83.5 ± 8.8 mm Hg, but neither 18a-GA, palmitoleic acid, nor carbenoxolone affected tone. Similarly, none of the agents affected vasorelaxation induced by the K + -channel activator levcromakalim. These results support the proposal that anandamide may act via activation of the Na + pump, but not via gap junction activation. Data are shown as mean ± S.E.M.

vasoconstriction in the spleen (Wagner et al., 2001). It is possible that synergistic interactions of cannabinoids with circulating and locally released vasocontractile mediators accounts for the pronounced, albeit transient, pressor response observed upon systemic administration of cannabinoids in anaesthetised animals. 3.3. Modulation of sympathetic neurotransmission mRNA for cannabinoid receptors has been detected in sympathetic and parasympathetic ganglia (Ishac et al., 1996; Buckley et al., 1998), and functional studies in vitro have confirmed that cannabinoid receptors can be expressed prejunctionally on peripheral autonomic nerves and act to inhibit neurotransmitter release. The classical studies demonstrating cannabinoid modulation of peripheral sympathetic neurotransmission have been carried out in the vas deferens and myenteric neurones in the intestine, where CB1 receptors mediate inhibition of neurotransmission (Pertwee et al., 1996a, 1996b). A greater understanding of the likely mechanisms involved has come from studies of neurones in the CNS, where CB1 receptors are highly expressed in specific brain regions, and in cells overexpressing the CB1 receptor. These have shown that the CB1 receptor, via Gi/o-protein coupling, inhibits N-type and P/Q-type Ca2 + channels and adenylyl cyclase, and activates inwardly rectifying K + channels (Felder et al., 1995, 1998; Mackie et al., 1993, 1995; Derkinderen et al., 1996;

Twitchell et al., 1997), which would lead to inhibition of neurotransmitter release. There is, however, conflicting evidence concerning the effects of cannabinoids on neurotransmission in the heart and blood vessels. Ishac et al. (1996) demonstrated anandamide- and D9-tetrahydrocannabinol-mediated inhibition of [3H]noradrenaline release in rat isolated atria that was blocked by the antagonist SR141716A, indicating an involvement of CB1 receptors. However, in rat and mouse atria and rat mesenteric blood vessels, Lay et al. (2000) have demonstrated that neither synthetic cannabinoid ligands nor anandamide influence sympathetic neurotransmission. On the other hand, pronounced inhibition of sympathetic neurotransmission by different chemical classes of cannabinoids has been reported in the rat mesenteric arterial bed (Ralevic & Kendall, 2000). These effects were not reversed by CB1 and CB2 receptor antagonists, indicating a possible involvement of a receptor distinct from CB1 or CB2. In conclusion, some, but not all, of these studies are in line with the proposal that peripheral prejunctional inhibition of sympathetic neurotransmission may be involved in the secondary hypotensive action of cannabinoids observed on systemic administration. (Fig. 4). 3.4. Modulation of sensory neurotransmission One of the most recent proposals to account for the vasodilatation in response to anandamide has been that it

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

197

Fig. 4. Diagram showing modulation of perivascular neurotransmission by cannabinoids in blood vessels. Cannabinoids may be released from the cell membrane of sympathetic and sensory nerves, together with classical and other novel neurotransmitters. Rapid uptake means that diffusion is restricted; hence, its principal action is likely to be as a prejunctional neuromodulator. Cannabinoids can act prejunctionally to inhibit sympathetic and sensory neurotransmitter release. Cannabinoid CB1 receptors are expressed on sympathetic nerves, but the cannabinoid receptor mediating inhibition of sensory neurotransmission by cannabinoids remains to be characterised. In pathophysiological situations, high levels of the endogenous cannabinoid anandamide (AEA), generated locally, for example, from macrophages in inflammation, may act at vanilloid VR1 receptors to activate sensory nerves, leading to release of sensory neurotransmitter and vasorelaxation. A1, A1 adenosine receptor; a1/2, a1/2-adrenoceptor; Ado, adenosine; cAMP, cyclic AMP; cGMP, cyclic GMP; NA, noradrenaline; NKA, neurokinin A; NPY, neuropeptide Y; P2X, purine P2X receptor; SP, substance P.

acts as a vanilloid agonist. Anandamide shares structural similarities with the vanilloid agonist olvanil. This prompted Zygmunt et al. (1999) to investigate the role of VRs in the vascular actions of anandamide. In this respect, they reported that vasorelaxation to anandamide (but not 2-AG, palmitoylethanolamide, or synthetic cannabinoid receptor agonists) was essentially abolished by depletion of the sensory nerves of calcitonin gene-related peptide (CGRP) by capsaicin in guinea-pig basilar, rat hepatic, and rat mesenteric arteries. In addition, the responses to anandamide were sensitive to the VR antagonist capsazepine and also to CGRP receptor antagonism with CGRP (8– 37). Clearly, anandamide can evoke the release of neurotransmitters from sensory nerves, leading to vasorelaxation. This conclusion was supported by the demonstration that anandamide is an agonist at the cloned rat VR1 (Zygmunt et al., 1999). This observation was later confirmed at the human VR1 (Smart et al., 2000). Similar observations have been made with the analogue of anandamide, methanandamide, which was also shown to cause capsaicin- and capsazepine-sensitive vasorelaxation in the rat mesenteric arterial bed and isolated mesenteric arteries (Ralevic et al., 2000b). However, in the same vascular bed, Harris et al. (2002) reported that vasorelaxation to anandamide was only partly sensitive to capsaicin pretreatment. Moreover, in the presence of NO synthase

blockade, vasorelaxation due to anandamide was insensitive to capsaicin pretreatment and, thus, does not occur exclusively via sensory nerves. In addition, relaxation to anandamide was augmented by AM404 (Fig. 1) (Harris et al., 1998), a compound with complex actions, including inhibition of cannabinoid transport and activation of VRs. Accordingly, the activation of sensory nerves by anandamide may only explain part of the actions of anandamide, and only under some circumstances. The fact that the hypotensive action of anandamide is absent in mice lacking CB1 receptors (Ledent et al., 1999) suggests that any action via VRs on sensory nerves is only of minor importance in the haemodynamic profile of systemically administered cannabinoids. There has been some debate about whether sensory nerves have a role in the in vivo vasodepressor action of anandamide (see Smart & Jerman, 2000; Szolcsa´nyi, 2000; Zygmunt et al., 2000b). An important point to emerge was that capsaicin, which mimics the triphasic haemodynamic response elicited by anandamide in anaesthetised animals, elicits only a vasopressor response in pithed rats. This indicates that activation of efferent function of sensory nerves is not involved in its hypotensive response, so it is hardly surprising that this is not involved in the depressor response to the less potent VR1 agonist anandamide. An implication is that ananda-

198

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

mide activation of VR1 receptors on sensory nerves, leading to vasodilatation in vitro, may be significant principally in the local control of vascular tone and/or in pathophysiological situations when high concentrations of cannabinoids are generated locally. Cannabinoid CB1 receptor mRNA has been localised in dorsal root ganglion (DRG) cells (Hohmann & Herkenham, 1999a; San˜udo-Pen˜a et al., 1999), and there is evidence of trafficking of CB1 receptors from the DRG to peripheral sensory nerves (Hohmann & Herkenham, 1999b). Anandamide has been shown to inhibit CGRP release via CB1 receptors in the skin of the rat hindpaw (Richardson et al., 1998). In addition, the synthetic cannabinoids HU210 and WIN-55212-2 have been shown to inhibit prejunctionally sensory neurotransmission in the rat mesenteric arterial bed (Ralevic & Kendall, 2001). HU210 and WIN-55212-2 do not activate VR1, excluding possible actions at VRs from this inhibitory effect. In addition, in experiments using rat cultured DRG neurones, we have found that HU210 inhibits capsaicin-induced Ca2 + influx in a manner sensitive to SR141716 (Millns et al., 2001). A very recent study has reported co-localisation of VR1 and CB1 receptors in DRG neurones (Ahluwalia et al., 2000), and Ross et al. (2001) have described the expression of both CB1 and CB2 receptors by DRG neurones. It appears, therefore, that cannabinoids, possibly via a CB1 receptor, can act as a functional brake on VR-mediated sensory nerve activation. There is still the question of why are essentially opposite effects, activation of the efferent function of sensory nerves via VR1 and inhibition of neurotransmitter release via cannabinoid receptors, elicited by the same agonist? In some respects, this is similar to the paradox that VR1 mediates vasorelaxation to cannabinoids in vitro, but not in vivo. In model cell systems transfected with human receptors, anandamide has very similar potencies for activation of CB1 (pEC50, 5.7 for inhibition of cyclic AMP formation) (Bonhaus et al., 1998) and VR1 receptors (pEC50, 5.9 for Ca2 + entry) (Smart et al., 2000). This calls into question the potential physiological role of anandamide with regard to sensory nerve control, given that it is difficult to conceive how a differential effect on CB1 and VR1 would be possible if the sensitivities of the receptors to the agonist are so similar. More subtle forms of differentiation, such as relative rates of receptor desensitisation in response to anandamide, might be relevant. Alternatively, the involvement of non-CB1 receptors and other endogenous cannabinoids might be important in sensory nerve control. DRG neurones also express CB2 receptors at which 2-AG is a full agonist, in contrast to anandamide, which is a weak partial agonist, (Gonsiorek et al., 2000). 2-AG, therefore, could be the ‘‘natural’’ agonist for CB2 and CB1 receptors, with anandamide acting as an endogenous VR1 agonist, as well as antagonising the effects of 2-AG at the CB2 receptor. This hypothesis is supported by the fact that 2-AG does not mimic the action of anandamide at VRs (Zygmunt et al., 1999) (Fig. 4).

4. Sites of action 4.1. Vascular cannabinoid receptors The ability of cannabinoids to cause vascular effects implies that the vasculature contains a molecular target. Evidence to date suggests that there may be vascular cannabinoid receptors, which may either fall into the classical CB1/ CB2 classification or represent a new subtype. As stated in Section 2.1, the biphasic hypotension in response to anandamide is absent in CB1 receptor knockout mice (Ledent et al., 1999). This clearly points to the involvement of the CB1 receptor. However, it should be noted that this does not conclusively identify the cannabinoid receptors as being on the vascular smooth muscle or being associated with neuronal tissue. The sensitivity of vasorelaxant responses to CB receptor antagonists has been controversial, with some studies indicating that the responses are opposed by SR141716A (Randall et al., 1996; White & Hiley, 1997) and others demonstrating that they are insensitive to this antagonist (Plane et al., 1997). The insensitivity to SR141716A might reflect non-cannabinoid receptor actions (Pratt et al., 1998; Chaytor et al., 1999), whilst Jarai et al. (1999) have suggested the presence of a novel vascular CB receptor. Using reverse transcriptase – polymerase chain reaction, the gene product encoding for CB1 receptors has been located in renal endothelial cells, mesenteric resistance arterioles, and cerebral microvessels, which is consistent with the expression of CB1 receptors in the vasculature (Deutsch et al., 1997; Darker et al., 1998; Randall et al., 1999). Using immunohistochemistry, antibody staining for CB1-like immunoreactivity was associated with both the endothelium and smooth muscle in rat mesenteric vessels, but not in the rat thoracic aorta (which does not relax to anandamide) (Randall et al., 1999). Others have also identified mRNA in human endothelial cells (Sugiura et al., 1998; Liu et al., 2000), and Liu et al. (2000) have identified CB1 receptor-binding sites by radioligand studies. Similarly, immunoreactivity to the CB1 receptor has been identified on human saphenous vein endothelial cells (Bilfinger et al., 1998). Cannabinoid CB1 receptors have also been localised to cat cerebral arterial smooth muscle (Gebremedhin et al., 1999). In this study, it was demonstrated that feline vascular smooth muscle contained CB1 receptors together with cDNA, showing very close homology to that associated with neuronal CB1 receptors. 4.2. Receptor-independent actions of cannabinoids Cannabinoids are highly lipophilic molecules, and readily penetrate the cell membrane. Indeed, it was not until specific cannabinoid receptors were cloned that conclusive evidence was provided that mechanisms other than alteration of the physiochemical properties of the cell membrane and/or an intracellular site could account for their vascular effects. Hence, where effects mediated by

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

cannabinoids are not selectively antagonised, the possibility that some of their vascular actions are mediated independently of cell surface cannabinoid receptors cannot be excluded.

5. Where are endocannabinoids produced in the vasculature? If endocannabinoids are to play a significant role in the regulation of vascular function, they must be produced close to their site of action because of rapid uptake and degradation. The membranes of cells in the blood vessels (endothelial and nerve cells) or activated blood cells are potential sources of endocannabinoids (Fig. 5). Our proposal that anandamide might be an endotheliumderived autacoid is consistent with the demonstration that cultured rat renal endothelial cells contain anandamide, together with synthase and amidase activities (Deutsch et al., 1997). In addition, cultured HUVEC release the endocannabinoid 2-AG on stimulation with a Ca2+ ionophore (Sugiura et al., 1998). Furthermore, Mechoulam et al. (1998) detected the release of 2-AG from rat aortae following stimulation with an endothelium-dependent relaxant. By analogy with the CNS, where anandamide is released from neurones following membrane depolarisation and Ca2 + influx, perivascular nerves may be a source of endocannabinoids. Indeed, perivascular sensory nerves have been proposed as a potential site of endocannabinoid production and release. In this respect, Ishioka and Bukoski (1999) demonstrated that sensory nerve-dependent Ca2 + -induced relaxation of rat mesenteric vessels was blocked by SR141716A, with the suggestion that an endocannabinoid was released

199

from the nerves and mediated the vasorelaxation. Moreover, in the rat isolated mesenteric arterial bed, SR141716A augmented sensory neurogenic vasorelaxation mediated by electrical stimulation, consistent with the idea that endocannabinoids may be released during depolarisation of the nerves (Ralevic & Kendall, 2000). In this study, SR141716A had no effect on the responses to exogenous CGRP, indicating that the released endocannabinoid might be acting prejunctionally as a neuromodulator. Neurogenic vasorelaxation mediated by electrical stimulation of the mesenteric arterial bed is virtually abolished by the CGRP receptor antagonist CGRP (8– 37) (Han et al., 1990). This is consistent with the suggestion that endocannabinoids released from sensory nerves do not have a direct postjunctional effect, but, rather, act as prejunctional neuromodulators, causing an inhibition of sensory neurotransmitter release and attenuation of the neurogenic relaxation response. The endocannabinoid-mediated sensory nervedependent vasorelaxation described by Ishioka and Bukoski (1999) was evoked using Ca2+ as a stimulant of sensory nerves, and under these conditions, extracellular endocannabinoids may have reached sufficiently high levels to evoke vasorelaxation postjunctionally. It should be noted that unlike classical neurotransmitters, anandamide is not stored in and released from vesicles, but rather it is thought to be synthesised on demand from the cell membrane by the phospholipase D-mediated hydrolysis of the phospholipid N-arachidonylphosphatidylethanolamine. This implies local neuromodulatory actions by the endocannabinoids, and, indeed, rapid degradation may make it more likely that endocannabinoids released from nerves would act prejunctionally rather than reaching specific targets postjunctionally. Whether there is a release of endocannabinoids from sympathetic perivascular nerves has not been specifically investigated yet.

Fig. 5. Diagram showing possible sources of endocannabinoid (anandamide, AEA) release and sites of uptake in blood vessels. AEA may be released on demand from the cell membrane of perivascular nerves to act as a neuromodulator via prejunctional cannabinoid receptors (CB) (and possibly via postjunctional receptors on the vascular smooth muscle to cause vasorelaxation). It may also be released from endothelial cells, possibly to act at CB receptors on the smooth muscle or to inhibit platelet aggregation. AEA may also be released from the cell membranes of blood-borne elements (platelets, leukocytes, macrophages) to act at CB receptors on the endothelium to cause vasorelaxation. Rapid uptake of AEA at the sites of release restricts the extent to which it can diffuse, implying that its actions are also restricted to the sites of release.

200

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

6. Endocannabinoids and pathophysiology Wagner et al. (1997) demonstrated in a rat model of haemorrhagic shock that activated macrophages release anandamide. In endotoxic shock, the synthesis of 2-AG in platelets is increased, and anandamide is only detectable in macrophages after exposure to lipopolysaccharide (Varga et al., 1998). In vitro, mouse J774 macrophages also release both 2-AG and anandamide, and participate in their degradation (Di Marzo et al., 1999). In patients with endotoxic shock, increases in plasma anandamide and 2AG have now been reported (Wang et al., 2001). These findings certainly point to the genesis of endocannabinoids in blood cells, which is enhanced in shock, and contributes toward the cardiovascular sequelae. In the context of septic shock, the induction of inducible NO synthase and excessive production of NO is widely implicated. Interestingly, Ross et al. (2000) demonstrated that the cannabinoid agonist WIN-55212, acting via CB2 receptors, actually inhibited lipopolysaccharide-induced NO release from macrophages. Liver cirrhosis in humans is associated with vasodilatation, and such patients are also endotoxaemic. In a rat model of liver cirrhosis, the apparent hypotension may be reversed by the CB1 receptor antagonist SR141716A. Furthermore, monocytes from cirrhotic patients have increased levels of anandamide, and there is an increase in the number of CB1 receptors on hepatic vascular endothelial cells from these patients (Ba´tkai et al., 2001). These findings suggest that endocannabinoids may contribute toward cardiovascular changes in cirrhosis, and may represent a novel therapeutic target. In terms of haemorrhagic shock, Wagner et al. (1997) demonstrated that the accompanying hypotension, in part due to macrophage-derived endocannabinoids, was reversed by the cannabinoid receptor antagonist SR141716A. Similarly, in endotoxic shock, the synthesis of 2-AG in platelets and anandamide in macrophages are increased (Varga et al., 1998). It is possible that the activated blood cells could also stimulate the release of endocannabinoids from the endothelium or other vascular sites, contributing further to the hypotensive response. The release of anandamide by central neurones under hypoxic conditions, leading to improved blood flow and protection against ischaemia, has also been advanced as a pathophysiological role for anandamide (Gebremedhin et al., 1999).

7. Concluding remarks Endocannabinoids exert potent and complex cardiovascular effects. Cannabinoids may inhibit peripheral neurotransmission by the inhibition of voltage-gated Ca 2+ channels and adenylyl cyclase, and activation of inwardly rectifying K + channels, but the precise mechanism(s) of cannabinoid-mediated vasorelaxation are uncertain. Whether

vasorelaxation has an endothelium-dependent component varies between species and between vascular sites. The responses may involve gap junctional communication, actions on the Na + pump, release of endothelial autacoids, inhibition of voltage-operated Ca2+ channels, or actions on VRs, leading to release of neurotransmitters from sensory nerves. Indeed, it is likely that there are several targets for the endocannabinoids, and experimentally (or pathologically) removing one target may be compensated for by action at another site. The fact that endocannabinoids have vascular effects raises the question of their relevance. As discussed in this review, there is emerging evidence pointing to their participation in shock, but what roles, if any, they play in normal physiology remains to be established.

Acknowledgements We thank the British Heart Foundation. D.H. holds an MRC Studentship and V.R. is a Research Fellow of the Royal Society.

References Ahluwalia, J., Urban, L., Capogna, M., Bevan, S., & Nagy, I. (2000). Cannabinoid 1 receptors are expressed in nociceptive primary sensory neurons. Neuroscience 100, 685 – 688. Ba´tkai, S., Ja´rat, Z., Wagner, J. A., Goparaju, S. K., Varga, K., Liu, J., Wang, L., Mirshahi, F., Khanolkar, A. D., Makriyannis, A., Urbaschek, R., Garcia, N., Sanyal, A. J., Jr., & Kunos, G. (2001). Endocannabinoids acting at vascular CB1 receptors mediate the vasodilated state in advanced liver cirrhosis. Nat Med 7, 827 – 832. Benowitz, N. L., & Jones, R. T. (1975). Cardiovascular effects of prolonged delta-9-tetrahydrocannabinol ingestion. Clin Pharmacol Ther 18, 287 – 297. Benowitz, N. L., Rosenberg, J., Rogers, W., Bachman, J., & Jones, R. T. (1979). Cardiovascular effects of intravenous delta-9-tetrahydrocannabinol: autonomic nervous mechanisms. Clin Pharmacol Ther 25, 440 – 446. Bilfinger, T. V., Salzet, M., Fimiani, C., Deutsch, A., Tramu, G., & Stefano, G. B. (1998). Pharmacological evidence for anandamide amidase in human cardiac vascular tissues. Int J Cardiology 64, S15 – S22. Bonhaus, D. W., Chang, L. K., Kwan, J., & Martin, G. R. (1998). Dual activation and inhibition of adenylyl cyclase by cannabinoid receptor agonists: evidence for agonist-specific trafficking of intracellular responses. J Pharmacol Exp Ther 287, 884 – 888. Buckley, N. E., Hansson, S., Harta, G., & Mezey, E. (1998). Expression of the CB1 and CB2 receptor messenger RNAs during embryonic development in the rat. Neuroscience 82, 1131 – 1149. Calignano, A., La Rana, G., Beltramo, M., Makriyannis, A., & Piomelli, D. (1997). Potentiation of anandamide hypotension by transport inhibitor, AM404. Eur J Pharmacol 337, R1 – R2. Chataigneau, T., Fe´le´tou, M., Thollon, C., Villeneuve, N., Vilaine, J.-P., Duhault, J., & Vanhoutte, P. M. (1998). Cannabinoid CB1 receptor and endothelium-dependent hyperpolarization in guinea-pig carotid, rat mesenteric and porcine coronary arteries. Br J Pharmacol 123, 968 – 974. Chaytor, A. T., Evans, W. H., & Griffith, T. M. (1998). Central role of heterocellular gap junctional communication in endothelium-dependent relaxations of rabbit arteries. J Physiol (Lond) 508, 561 – 573.

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202 Chaytor, A. T., Martin, P. E. M., Evans, W. H., Randall, M. D., & Griffith, T. M. (1999). The endothelial component of cannabinoid-induced relaxation in rabbit mesenteric artery depends on gap junctional communication. J Physiol (Lond) 520, 539 – 550. Chen, Y., McCarron, R. M., Ohara, Y., Bembry, J., Azzam, N., Lenz, F. A., Shohmani, E., Mechoulam, R., & Spatz, M. (2000). Human brain capillary endothelium-2-arachidonoglycerol (endocannabinoid) interacts with endothelin-1. Circ Res 87, 323 – 327. Darker, I. T., Millns, P. J., Selbie, L., Randall, M. D., S-Baxter, G., & Kendall, D. A. (1998). Cannabinoid (CB1) receptor expression is associated with mesenteric resistance vessels but not thoracic aorta in the rat. Br J Pharmacol 125, 95P. Derkinderen, P., Toutant, M., Burgaya, F., Le Bert, M., Siciliano, J. C., De Franciscis, V., Gelman, M., & Girault, J.-A. (1996). Regulation of a neuronal form of focal adhesion kinase by anandamide. Science 273, 1719 – 1722. Deutsch, D. G., Goligorsky, M. S., Schmid, P. G., Krebsbach, R. J., Schmid, H. H. O., Das, S. K., Dey, S. K., Arreaza, G., Thorup, C., Stefano, G., & Moore, L. C. (1997). Production and physiological actions of anandamide in the vasculature of the rat kidney. J Clin Invest 100, 1538 – 1546. Dewey, W. L. (1986). Cannabinoid pharmacology. Pharmacol Rev 38, 151 – 178. Di Marzo, V., Bisogno, T., DePetrocellis, L., Melck, D., Orlando, P., Wagner, J. A., & Kunos, G. (1999). Biosynthesis and inactivation of the endocannabinoid 2-arachidonoylglycerol in circulating and tumoral macrophages. Eur J Biochem 264, 258 – 267. Ellis, E. F., Moore, S. F., & Willoughby, K. A. (1995). Anandamide and D9THC dilation of cerebral arterioles is blocked by indomethacin. Am J Physiol (Heart Circ) 38, H1859 – H1864. Felder, C. C., Joyce, K. E., Briley, E. M., Mansouri, J., Mackie, K., Blond, O., Lai, Y., Ma, A. L., & Mitchell, R. L. (1995). Comparison of the pharmacology and signal transduction of the human cannabinoid CB1 and CB2 receptors. Mol Pharmacol 48, 443 – 450. Felder, C. C., Joyce, K. E., Briley, E. M., Glass, M., Mackie, K. P., Fahey, K. J., Cullinan, G. J., Hunden, D. C., Johnson, D. W., Chaney, M. O., Koppel, G. A., & Brownstein, M. (1998). LY320135, a novel cannabinoid CB1 receptor antagonist, unmasks coupling of the CB1 receptor to stimulation of cAMP accumulation. J Pharmacol Exp Ther 284, 291 – 297. Fleming, I., Schermer, B., Popp, R., & Busse, R. (1999). Inhibition of the production of endothelium-derived hyperpolarizing factor by cannabinoid receptor agonists. Br J Pharmacol 126, 949 – 960. Gardiner, S. M., March, J. E., Kemp, P. A., & Bennett, T. (2001). Regional haemodynamic responses to the cannabinoid agonist, WIN 55212-2, in conscious, normotensive rats, and in hypertensive, transgenic rats. Br J Pharmacol 133, 445 – 453. Garland, C. J., Plane, F., Kemp, B. K., & Cocks, T. M. (1995). Endothelium-dependent hyperpolarization: a role in the control of vascular tone. Trends Pharmacol Sci 16, 23 – 30. Gebremedhin, D., Lange, A. R., Campbell, W. B., Hillard, C. J., & Harder, D. R. (1999). Cannabinoid CB1 receptor of cat cerebral arterial muscle functions to inhibit L-type Ca2 + channel current. Am J Physiol (Heart Circ) 45, H2085 – H2093. Gonsiorek, W., Lunn, C., Fan, X., Narula, S., Lundell, D., & Hipkin, R. W. (2000). Endocannabinoid 2-arachidonyl glycerol is a full agonist through human type 2 cannabinoid receptor: antagonism by anandamide. Mol Pharmacol 57, 1045 – 1050. Han, S.-P., Naes, L., & Westfall, T. C. (1990). Inhibition of periarterial nerve stimulation-induced vasodilation of the mesenteric arterial bed by CGRP (8 – 37) and CGRP receptor desensitization. Biochem Biophys Res 168, 786 – 791. Harris, D., Kendall, D. A., & Randall, M. D. (1998). Effects of AM404, a cannabinoid reuptake inhibitor, on EDHF-mediated relaxations in the rat isolated mesentery. Br J Pharmacol 125, 15P. Harris, D., Martin, P. E. M., Evans, W. H., Kendall, D. A., Griffith, T. M., & Randall, M. D. (2000). Role of gap junctions in endothelium-derived

201

hyperpolarizing factor responses and mechanisms of K + -relaxation. Eur J Pharmacol 401, 119 – 128. Harris, D., McCulloch, I. A., Kendall, D. A., & Randall, M. D. (2002). Characterization of vasorelaxant responses to anandamide in the rat mesenteric arterial bed. J Physiol (Lond) 539, 893 – 903. Himmi, T., Perrin, J., Ouzzani, T. E., & Orsini, J.-C. (1998). Neuronal responses to cannabinoid receptor ligands in the solitary tract nucleus. Eur J Pharmacol 359, 49 – 54. Hohmann, A. G., & Herkenham, M. (1999a). Localization of central cannabinoid CB1 receptor messenger RNA in neuronal subpopulations of rat dorsal root ganglia: a double-label in situ hybridization study. Neuroscience 90, 923 – 931. Hohmann, A. G., & Herkenham, M. (1999b). Cannabinoid receptors undergo axonal flow in sensory nerves. Neuroscience 92, 1171 – 1175. Holland, M., Challis, R. A., Standen, N. B., & Boyle, J. P. (1999). Cannabinoid CB1 receptors fail to cause relaxation, but couple via Gi/Go to inhibition of adenylyl cyclase in carotid artery smooth muscle. Br J Pharmacol 128, 597 – 604. Ishac, E. J. N., Jiang, L., Lake, K. D., Varga, K., Abood, M. E., & Kunos, G. (1996). Inhibition of exocytotic noradrenaline release by presynaptic cannabinoid CB1 receptors on peripheral sympathetic nerves. Br J Pharmacol 118, 2023 – 2028. Ishioka, N., & Bukoski, R. D. (1999). A role for N-arachidonylethanolamine (anandamide) as the mediator of sensory nerve-dependent Ca2 + induced relaxation. J Pharmacol Exp Ther 289, 245 – 250. Jarai, Z., Wagner, J. A., Varga, K., Lake, K., Compton, D. R., Martin, B. R., Zimmer, A. M., Bonner, T. I., Buckley, N. E., Mezey, E., Razdan, R. K., Zimmer, A., & Razdan, G. (1999). Cannabinoid-induced mesenteric vasodilation through an endothelial site distinct from CB1 or CB2 receptors. Proc Natl Acad Sci USA 96, 14136 – 14141. Jarai, Z., Wagner, J. A., Goparaju, S. K., Wang, L., Razadan, R. K., Sugiura, T., Zimmer, A. M., Bonner, T. I., Zimmer, A., & Kunos, G. (2000). Cardiovascular effects of 2-arachidonoyl glycerol in anaesthetized mice. Hypertension 35, 679 – 684. Lake, K. D., Martin, B. R., Kunos, G., & Varga, K. (1997). Cardiovascular effects of anandamide in anaesthetized and conscious normotensive rats. Hypertension 29, 1204 – 1210. Lay, L., Angus, J. A., & Wright, C. E. (2000). Pharmacological characterisation of cannabinoid CB1 receptors in the rat and mouse. Eur J Pharmacol 391, 151 – 161. Ledent, C., Valverde, O., Cossu, C., Petitet, F., Aubert, L. F., Beslot, F., Bohme, G. A., Imperato, A., Pedrazzini, T., Roques, B. P., Vassart, G., Fratta, W., & Parmentier, M. (1999). Unresponsiveness to cannabinoids and reduced additive effects of opiates in CB1 receptor knockout mice. Science 283, 401 – 404. Liu, J., Gao, B., Mirshahi, F., Sanyal, A. J., Khanolkar, A. D., Makriyannis, A., & Kunos, G. (2000). Functional CB1 cannabinoid receptors in human vascular endothelial cells. Biochem J 346, 835 – 840. Maccarrone, M., Bari, M., Lorenzon, T., Bisogno, T., Di Marzo, V., & Finazzi-Agro, A. (2000). Anandamide uptake by human endothelial cells and its regulation by nitric oxide. J Biol Chem 275, 13484 – 13492. Mackie, K., Devane, W. A., & Hille, B. (1993). Anandamide, an endogenous cannabinoid, inhibits calcium currents as a partial agonist in N18 neuroblastoma cells. Mol Pharmacol 44, 498 – 503. Mackie, K., Lai, Y., Westenbroek, R., & Mitchell, R. (1995). Cannabinoids activate an inwardly rectifying potassium conductance and inhibit Q-type calcium currents in AtT20 cells transfected with rat brain cannabinoid receptor. J Neurosci 15, 6552 – 6561. Mechoulam, R., Fride, E., Ben-Shabat, S., Meiri, U., & Horowitz, M. (1998). Carbachol, an acetylcholine receptor agonist, enhances production in the rat aorta of 2-arachidonoyl glycerol, a hypotensive endocannabinoid. Eur J Pharmacol 362, R1 – R3. Millns, P., Chapman, V., & Kendall, D. A. (2001). Cannabinoid inhibition of capsaicin-induced calcium response in rat dorsal root ganglion cells. Br J Pharmacol 132, 969 – 971. Mombouli, J.-V., & Vanhoutte, P. M. (1997). Endothelium-derived hyper-

202

M.D. Randall et al. / Pharmacology & Therapeutics 95 (2002) 191–202

polarizing factor(s): updating the unknown. Trends Pharmacol Sci 18, 252 – 256. Niederhoffer, N., & Szabo, B. (1999). Effect of the cannabinoid receptor agonist WIN55212-2 on sympathetic cardiovascular regulation. Br J Pharmacol 126, 457 – 466. Pertwee, R. G., Fernando, S. R., Nash, J. E., & Coutts, A. A. (1996a). Further evidence for the presence of cannabinoid CB1 receptors in guinea pig small intestine. Br J Pharmacol 118, 2199 – 2205. Pertwee, R. G., Joe-Adigwe, G., & Hawksworth, G. M. (1996b). Further evidence for the presence of cannabinoid CB1 receptors in mouse vas deferens. Eur J Pharmacol 296, 169 – 172. Plane, F., Holland, M., Waldron, G. J., Garland, C. J., & Boyle, J. P. (1997). Evidence that anandamide and EDHF act via different mechanisms in rat isolated mesenteric arteries. Br J Pharmacol 121, 1509 – 1512. Pratt, P. F., Hillard, C. J., Edgemond, W. S., & Campbell, W. B. (1998). N-arachidonylethanolamide relaxation of bovine coronary artery is not mediated by CB1 cannabinoid receptor. Am J Physiol (Heart Circ) 43, H375 – H381. Ralevic, V., & Kendall, D. A. (2000). Cannabinoid modulation of sympathetic neurotransmission in the rat isolated mesenteric arterial bed. J Auton Nerv Syst 82, 76. Ralevic, V., & Kendall, D. A. (2001). Cannabinoid inhibition of capsaicinsensitive sensory neurotransmission in the rat isolated mesenteric arterial bed. Eur J Pharmacol 418, 117 – 125. Ralevic, V., Kendall, D. A., Bennett, T., & Gardiner, S. M. (2000a). Comparison of the effects of anandamide and methanandamide in bioassays. Br J Pharmacol 129, 141P. Ralevic, V., Kendall, D. A., Randall, M. D., Zygmunt, P. M., Movahed, P., & Ho¨gestatt, E. D. (2000b). Vanilloid receptors on capsaicin-sensitive nerves mediate relaxation to methanandamide in the rat isolated mesenteric bed. Br J Pharmacol 130, 1483 – 1488. Randall, M. D., & Kendall, D. A. (1997). The involvement of an endogenous cannabinoid in EDHF-mediated vasorelaxation in the rat coronary vasculature. Eur J Pharmacol 335, 205 – 209. Randall, M. D., & Kendall, D. A. (1998). Anandamide and endotheliumderived hyperpolarizing factor act via a common vasorelaxant mechanism in rat mesentery. Eur J Pharmacol 346, 51 – 53. Randall, M. D., Alexander, S. P. H., Bennett, T., Boyd, E. A., Fry, J. R., Gardiner, S. M., Kemp, P. A., McCulloch, A. I., & Kendall, D. A. (1996). An endogenous cannabinoid as an endothelium-derived vasorelaxant. Biochem Biophys Res Commun 229, 114 – 120. Randall, M. D., McCulloch, A. I., & Kendall, D. A. (1997). Comparative pharmacology of endothelium-derived hyperpolarizing factor and anandamide in rat isolated mesentery. Eur J Pharmacol 333, 191 – 197. Randall, M. D., Harris, D., Darker, I. T., Millns, P. J., & Kendall, D. A. (1999). Endocannabinoids: endothelium-derived vasodilators. In P. M. Vanhoutte (Ed.), Endothelium-Dependent Hyperpolarizations (pp. 149 – 155). Amsterdam: Harcourt. Richardson, J. D., Kilo, S., & Hargreaves, K. M. (1998). Cannabinoids reduce hyperalgesia and inflammation via interaction with peripheral CB1 receptors. Pain 75, 111 – 119. Ross, R. A., Brockie, H. C., & Pertwee, R. G. (2000). Inhibition of nitric oxide production in RAW264.7 macrophages by cannabinoids and palmitoylethanolamine. Eur J Pharmacol 401, 121 – 130. Ross, R. A., Coutts, A. A., McFarlane, S. M., Anavi-Goffer, S., Irving, A. J., Pertwee, R. G., MacEwan, D. J., & Scott, R. H. (2001). Actions of cannabinoid receptor ligands on rat cultured sensory neurones: implications for antinociception. Neuropharmacology 40, 221 – 232. San˜udo-Pen˜a, M. C., Strangman, N. M., Mackie, K., Walker, J. M., & Tsou, K. (1999). CB1 receptor localization in rat spinal cord and roots, dorsal root ganglion, and peripheral nerve. Acta Pharmacol Sin 20, 1115 – 1120. Smart, D., & Jerman, J. C. (2000). Anandamide: an endogenous activator of the vanilloid receptor. Trends Pharmacol Sci 21, 134. Smart, D., Gunthorpe, M. J., Jerman, J. C., Nasir, S., Gray, J., Muir, A. I., Chambers, J. K., Randall, A. D., & Davis, J. B. (2000). The endogenous

lipid anandamide is a full agonist at the human vanilloid receptor (hVR1). Br J Pharmacol 129, 227 – 230. Stark, P., & Dews, P. B. (1980). Cannabinoids. II. Cardiovascular effects. J Pharmacol Exp Ther 214, 131 – 138. Stein, E. A., Fuller, S. A., Edgemond, W. S., & Campbell, W. B. (1996). Physiological and behavioural effects of the endogenous cannabinoid, arachidonylethanolamine (anandamide), in the rat. Br J Pharmacol 119, 107 – 114. Sugiura, T., Kodaka, T., Nakane, S., Kishimoto, S., Kondo, S., & Waku, K. (1998). Detection of an endogenous cannabimimetic molecule, 2-arachidonoylglycerol, and cannabinoid CB1 receptor mRNA in human vascular cells: is 2-arachidonoylglycerol a possible vasomodulator. Biochem Biophys Res Commun 243, 838 – 843. Szolcsa´nyi, J. (2000). Are cannabinoids endogenous ligands for the VR1 capsaicin receptor. Trends Pharmacol Sci 21, 41 – 42. Tsou, K., Brown, S., Sanudo-Pena, M. C., Mackie, K., & Walker, J. M. (1998). Immunohistochemical distribution of cannabinoid CB1 receptors in the rat central nervous system. Neuroscience 83, 393 – 411. Twitchell, W., Brown, S., & Mackie, K. (1997). Cannabinoids inhibit Nand P/Q-type calcium channels in cultured rat hippocampal neurons. J Neurophysiol 78, 43 – 50. Varga, K., Lake, K., Martin, B. R., & Kunos, G. (1995). Novel antagonist implicates CB1 cannabinoid receptor in the hypotensive action of anandamide. Eur J Pharmacol 278, 279 – 283. Varga, K., Lake, K. D., Huangfu, D., Guyenet, P. G., & Kunos, G. (1996). Mechanism of the hypotensive action of anandamide in anaesthetised rats. Hypertension 28, 682 – 686. Varga, K., Wagner, J. A., Bridgen, D. T., & Kunos, G. (1998). Platelet- and macrophage-derived endogenous cannabinoids are involved in endotoxin-induced hypotension. FASEB J 12, 1035 – 1044. Vidrio, H., Sa´nchez-Salvatori, M. A., & Medina, M. (1996). Cardiovascular effects of ( )-11-OH-D8-tetrahydrocannabinol-dimethylheptyl in rats. J Cardiovasc Pharmacol 28, 332 – 336. Wagner, J. A., Varga, K., Ellis, E. F., Rzigalinski, B. A., Martin, B. R., & Kunos, G. (1997). Activation of peripheral CB1 cannabinoid receptors in haemorrhagic shock. Nature 390, 518 – 521. Wagner, J. A., Varga, K., Jarai, Z., & Kunos, G. (1999). Mesenteric vasodilation mediated by endothelial anandamide receptors. Hypertension 33, 429 – 434. Wagner, J. A., Ja´rai, Z., Ba´tkai, S., & Kunos, G. (2001). Haemodynamic effects of cannabinoids: coronary and cerebral vasodilation mediated by cannabinoid CB1 receptors. Eur J Pharmacol 423, 203 – 210. Wang, Y., Liu, Y., Hashiguchi, T., Kitajima, I., Yamakuchi, M., Shimizu, H., Matsuo, S., Imaizuma, H., & Maruyama, I. (2001). Simultaneous measurement of anandamide and 2-arachidonoylglycerol by polymyxin B-selective adsorption and subsequent high-performance liquid chromatography analysis: increase in endogenous cannabinoids in the sera of patients with endotoxic shock. Anal Biochem 294, 73 – 82. White, R., & Hiley, C. R. (1997). A comparison of EDHF-mediated and anandamide-induced relaxations in the rat isolated mesenteric artery. Br J Pharmacol 122, 1573 – 1584. White, R., & Hiley, C. R. (1998). The actions of some cannabinoid receptor ligands in the rat isolated mesenteric artery. Br J Pharmacol 125, 533 – 541. Zygmunt, P. M., Ho¨gestatt, E. D., Waldeck, K., Edwards, G., Kirkup, A. J., & Weston, A. H. (1997). Studies on the effects of anandamide in the rat hepatic artery. Br J Pharmacol 122, 1679 – 1686. Zygmunt, P. M., Petersson, J., Anderssson, D. A., Chuang, H.-h., Sørga˚rd, M., Di Marzo, V., Julius, D., & Ho¨gesta¨tt, E. D. (1999). Vanilloid receptors on sensory nerves mediate the vasodilator action of anandamide. Nature 400, 452 – 457. Zygmunt, P. M., Chuang, H., Movahed, P., Julius, D., & Ho¨gesta¨tt, E. D. (2000a). The anandamide transport inhibitor AM404 activates vanilloid receptors. Eur J Pharmacol 396, 39 – 42. Zygmunt, P. M., Julius, D., Di Marzo, V., & Ho¨gesta¨tt, E. D. (2000b). Anandamide—the other side of the coin. Trends Pharmacol Sci 21, 43 – 44.